The Learning Curve for Intraoperative Echocardiography During Congenital Heart Surgery Ross M. Ungerleider, MD, William J. Greeley, MD, Ronald J. Kanter, MD, and Joseph A. Kisslo, MD Department of Surgery, Duke University Medical Center, Durham, North Carolina

Our group has previously reported a large prospective experience with the use of intraoperative echocardiography with Doppler color-flow imaging (IE-DCFI)during the repair of congenital heart defects. We have now performed IE-DCFI in 621 patients and have observed a major change in the impact of this technology, which has stabilized during our most recent experience (the last 207 patients). To evaluate the surgical learning curve with IE-DCFI, we divided patients into three groups: group 1, patients 1through 207; group 2, patients 208 through 414; and group 3, patients 415 through 621. There were no major differences between groups with respect to age or disease entities. The average time needed to perform an IE-DCFI examination decreased from 3.75 2 1.77 minutes in group 1 to 3.35 2 1.52 minutes in group 2 and has remained stable. The number of patients requiring revisions in the operating room (based on IE-DCFI findings)

decreased from 17 (8%)in group 1 to 7 (3%)in group 2 to 5 (2%) in group 3. Furthermore, revisions were 100% successful in correcting the problem in groups 2 and 3, whereas 18% of group 1 patients left the operating room

I

ital heart surgeon can learn how to obtain and interpret his or her own images and make decisions regarding the quality of repair that will enhance the patient’s long-term outcome. Although the refinement of these skills probably occurs over the course of a career, this report describes the learning curve that should be expected and accepted by a surgeon who wishes to eventually become competent with this technology.

ntraoperative echocardiography is becoming an increasingly used technology for the evaluation of repair of congenital heart defects, but familiarity with its use by surgeons remains the single most important limitation to its application in any given institution. Although pediatric cardiologists and anesthesiologists are helpful in obtaining and interpreting echocardiographic images, it is the surgeon who has the most cogent understanding of the anatomy and the most lucid mental image of the complex intracardiac spatial relationships created by the reconstruction. Only the surgeon can relate ultrasound images to his or her own personal concerns with respect to specific areas of the repair. In this sense, it is the surgeon who should be most capable of interpreting the intraoperative echo in relation to his or her knowledge of how the repair proceeded and attach appropriate relevance to the echo findings. Nevertheless, surgeons are not ordinarily trained in echocardiography, and the advantages that surgeons possess for performing clinically relevant and precise interpretation of intraoperative echocardiograms is more often than not neutralized by lack of experience with and understanding of this technology. However, with assistance from cardiologists and anesthesiologists, a congenPresented at the Thirty-eighth Annual Meeting of the Southern Thoracic Surgical Association, Orlando, FL, Nov 7-9, 1991. Address reprint requests to Dr Ungerleider, Department of Surgery, Duke University Medical Center, Box 3178, Durham, NC 27710.

0 1992 by The Society of Thoracic Surgeons

with persistent residual defects by IE-DCFI. Surgeons can acquire the ability to interpret the results of IE-DCFI themselves and use it to enhance their operative repair of congenital heart defects, but this requires an experience of at least 200 cases. Once skills are refined, the technique will identify 2% to 3% of patients who would benefit from operative revision before leaving the operating room and will direct that revision. As we learned to more accurately evaluate IE-DCFI data, the likelihood of a good outcome for patients leaving the operating room without IE-DCFI-identified problems improved from 88% in group 1 to 95% in group 2 to 97% in group 3. (Ann Thorac Surg 1992;54:691-8)

Material and Methods Intraoperative echocardiograms were obtained before and after repair from 621 patients undergoing repair of congenital heart defects on cardiopulmonary bypass through a median sternotomy. The majority of these examinations were performed using an epicardial technique previously described by our group [l].In some patients, transesophageal echocardiography was also performed depending on the nature of the defect. A cardiology colleague scrubbed and assisted with the epicardial imaging during the first 300 cases. Nevertheless, the surgeon attempted to obtain views in all patients, and ultimately (the last half of the series) a cardiologist scrubbed to assist the surgeon in obtaining images only if specifically requested to help evaluate a particular problem. All images were recorded on high-fidelity tape and were later reviewed by cardiologists experienced in echocardiography. However, decisions made at the time of operation were based on 0003-4975/92/$5.OO

692

UNGERLEIDER ET AL ECHO FOR CONGENITAL HEART SURGERY

Ann Thorac Surg 1992;54:691-8

Table 1 . Outcome Relating to Final Echo Results" Postrepair Grade

No. of Patientsb

Reoperation

Echo perfect Acceptable residual Unacceptable residual

375 (60) 204 (33) 39 (6)

2 (0.5) 20 (10) 13 (33)

a Numbers in parentheses are percentages. entire series of 621 patients.

Percentages are based on

real-time information and on the operating team's impression of these data. Although the results of prospective information generated from the first half of the patients in this study have been previously reported, this particular study analyzes the data from 621 patients in a retrospective fashion to identify (1)the incidence of repair revision in the operating room based on information generated by echocardiography performed after bypass and (2) the ability of the surgeon to properly identify an acceptable repair (by evaluating the need for reoperation for patients over the entire course of the study). Patients were divided into three groups based on chronological appearance in the series. Patients were entered if they met the criteria of undergoing correction of a congenital cardiac defect through a median sternotomy on cardiopulmonary bypass. The first entry (patient 1) was in March 1987 and the final entry, July 1991. For the purposes of retrospective analysis, patients 1 through 207 were labeled as group 1 (the early experience), patients 208 through 414 as group 2 (the intermediate experience), and patients 415 through 621 as group 3 (the most recent experience). Patients ranged in age from 1 day to 59 years, with 240 patients (39%) less than 1 year of age and 417 (67%) less than 3 years of age. The smallest patient in the series weighed 1,700 g and underwent closure of a ventricular septa1 defect. Patients were followed up with respect to outcome after operation. A bad outcome was considered to be reoperation (for a residual defect) or death. Reoperation that occurs as the natural history of an operative procedure (eg, conduit replacement after a Rastelli procedure) was

Table 2 . Outcome Relating to Final Echo Results" Postrepair Grade

Echo-perfect Acceptable residual Unacceptable residual

Group

No. of Patientsb

Reoperation

1 2 3 1 2 3 1 2 3

138 (67) 124 (60) 113 (55) 47 (23) 68 (33) 89 (43) 20 (10) 14 (7) 5 (2)

2 (1.4) 0 0 13 (28) 4 (6) 3 (3) 8 (40) 4 (29) 1(20)

Numbers in parentheses are percentages. number of patients in each group.

a

Percentages are based on

not coded as a bad outcome, but the need for valve replacement after valve repair was considered an operative failure, and the reoperation was coded as such. Follow-up has been complete as of November 1, 1991. This allows a full 3 months of follow-up since the latest entry in the series. It has been our experience through the evolution of this series that the majority of infants who require reoperation for failure of a previous procedure do so within the first 6 to 8 weeks after operation.

Results All three groups were equivalent with respect to major cardiac defects, male to female ratio, and age distribution. With experience, it was found that the average time for an intraoperative epicardial echo examination fell from 3.75 -C 1.77 minutes in group 1to 3.35 k 1.51 minutes in group 2 and has remained relatively stable over the recent months. The time for an examination is initially dependent on surgeon familiarity and skill with the technique but eventually is related more to the nature of the cardiac defect, the information needed, and the echogenicity of the patient. Results were graded as follows: (1) "echo-perfect" (no discernible residual defect, with good ventricular function and a repair that met or exceelded all expectations); (2) "acceptable" residual defect (a repair that disclosed discernible residual defects that were not thought to be major and that could be consistent with a good outcome); or (3) "unacceptable" residual defect (signifying a suboptimal result; often these patients were not considered to be in a condition that allowed revision of the repair, or it was not thought that the repair could be performed any better). Table 1 shows the correlation of final echo results as they were coded at the completion of each procedure with the incidence of reoperation during follow-up. Table 2 separates these results by patient groups. It is notable (Table 2) that with increasing experience and skill (groups 2 and 3) in obtaining intraoperative images, surgeons are more likely to find residual defects and are less likely to code the repair as echo-perfect. Nevertheless, the understanding of which results truly are acceptable correlates with increasing experience, and the incidence of patients requiring reoperation in this category becomes a very appropriate 3%to 6%. Furthermore, with improved skills (as echo teaches the surgeon how and where to avoid technical errors), the incidence of i3n unacceptable technical result becomes less common. During the course of the study, records were maintained on all patients who had revision of the repair in the operating room based on echo information alone. In the initial phases of the study, oxyg;en saturation information or green dye curves were obtained to verify the importance of the echo findings. However, the correlation was not always accurate, and it was found that with improved understanding of the echocardiiographicimages, the surgeons were able to obtain a much more sensitive impression of the repair quality as well as specific information regarding any regions of difficulty [2]. Table 3 lists the number of revisions performed in each group on patients

UNGERLEIDERET AL ECHO FOR CONGENITAL HEART SURGERY

Ann Thorac Surg 1992;54691-8

whose repairs were suboptimal and notes whether or not the revision was successful in producing a good operative result. Intraoperative revision was "common" in group 1 and was not always successful in producing a good technical result. With experience, it was found that the need of revision of the operation decreased but was still necessary in 2% to 3% of patients (groups 2 and 3). In these latter groups, echo invariably guided the surgeon efficiently and in a specific manner to achieve a technically good result. The surgeon's impression of the postrepair echo results and the patient's long-term outcome were compared. Death or the need for reoperation during follow-up was considered a bad outcome. Any concern with the postrepair echo was noted, including residual defects, severe ventricular dysfunction, or serious remaining abnormality of the reconstructed anatomy. Patients for whom there were no concerns regarding the postrepair echo (473 patients, 76%) had an excellent likelihood (92%, 434 patients) of a good long-term outcome, but this likelihood fell to 41% (61 patients) for patients whose postrepair echo demonstrated any reason for concern (24%, 148 patients). Table 4 reexamines these data with respect to the study groups and demonstrates that with experience (groups 2 and 3), the ability of the surgeon to correlate the postrepair echo with the likelihood of a good long-term outcome approaches 95% to 97%. Likewise, based on our current experience, the likelihood of a good outcome when there is a reason for concern regarding the postrepair echo falls to only 24%.

Comment Intraoperative echocardiography has been reported by us [2, 31 and several others [4-91 to be beneficial during the repair of congenital heart defects. The purpose of this report is not to revalidate this concept, although this conclusion could again be clearly drawn from the substantial documentation that has accumulated from this large series of 621 patients. Rather, it has been our interest to display these data in a retrospective fashion to demonstrate what became intuitively obvious to us throughout the course of the study-that there is a synergistic relationship between a surgeon's experience and the contributions of intraoperative echocardiography. A learning curve is a subjective impression acquired by a surgeon that he or she has gained important experience dealing with a set of problems and that this has contributed to an improvement in results during the involved time frame. There has been a clear improvement in surgical results within the time frame of this study. We have not reoperated on a single patient for residual ventricular septa1 defect since 1988, and we are thoroughly and enthusiastically convinced that information obtained from intraoperative echocardiography has resulted in the more rapid evolution of proper surgical techniques with immediate reinforcement of the most effective methods for repairing a variety of defects. The fact that a surgical team can learn to use intraoperative echocardiography to make relevant and appropriate

693

Table 3. Revisions in Operating Room for 621 Patients' Patient Group

Revisions Based on Echo

Good Result 14 (82) 7 (100) 5 (1W

1 2 3 a

Numbers in parentheses are percentages.

decisions is underscored by the fact that a cardiologst was not routinely present during the operations on the last 300 patients in this series. It was during this same period that the most dependable correlation was being obtained between the intraoperative echo and the ability to associate the findings with patient outcome. For a surgical team to acquire the necessary skills and experience with this technology, it is important for them to have adequate and committed support and supervision during the early phases of learning. In our case, we were fortunate to have the consistent and dedicated involvement of an experienced echocardiographer (J.A.K.) for every procedure performed in group 1 and many of the procedures in group 2. Surgeons can be taught how to manipulate an epicardial transducer and to interrogate intracardiac anatomy in a thorough and accurate fashion. The process of placing a transducer on the heart is not difficult, but the ability to interpret the images generated and the manipulation of the probe to display unusual but important views [lo, 111 are acquired skills that must be taught and learned over time. Comparison of Tables 1 and 2 demonstrates the impact of experience. When evaluated as a complete series of 621 patients, it appears that leaving the operating room with an acceptable residual defect, as demonstrated by postrepair echocardiography, is a risk factor for reoperation. If this were truly the case, then selection of the term acceptable for these residual defects would be inappropriate. However, reevaluation of our evolution of the term acceptable, when analyzed by looking at the early versus the later patients in the series, demonstrates that with experience, the surgical team is able to learn how to better correlate an acceptable residual defect with the reality of long-term outcome. It is interesting that as experience with the technology increased, it was less likely for us to find an echo-perfect repair despite the fact that the nature

Table 4. Outcome Relating to Echo Results" Postrepair Echo No. of Results Group Patientsb No concerns

Reason for concern

1 2 3 1 2 3

Numbers in parentheses are percentages. number of patients in each group (n = 207).

a

140 (68) 155 (75) 178 (86) 67 (32) 52 (25) 29 (14)

Good

Outcome 115 (82) 147 (95) 172 (97) 37 (55) 17 (33) 7 (24)

Percentages are based on

694

UNGERLEIDERET AL ECHO FOR CONGENITAL HEART SURGERY

of operations between groups was essentially the same. It is also interesting that there were two reoperations in group 1 for patients who presumably left the operating room with an echo-perfect repair. More likely, with increased skill in using this technology, the surgical team became better capable of recognizing, and recording as such, echo-imperfect repairs for reasons that might have been unappreciated in the early part of the series. Correlating with this is the increase in acceptable residual defects in group 3 compared with group 1. Again, this is probably a reflection of more experience in interpretation of the postrepair echo. Most importantly, however, there was a better understanding of what truly constituted an acceptable residual defect by the time our series had reached 200 patients so that the incidence of reoperation in this group of patients became more consistent with the term acceptable. Although the definition of acceptable is relatively subjective, it is important to emphasize the fact that with time and experience, a surgeon who pays attention to how the postrepair echo findings compare with patient outcome will eventually be able to have an increasingly clear understanding of what constitutes an acceptable residual defect after the repairs. A learning curve comprises subjective interpretation of experience. In the case of intraoperative echo, learning results in (1) better recognition of small residual defects, (2) more accurate prediction of which defects by echo are acceptable (as supported by correlation with acceptable outcome for the patient), and (3) fewer unacceptable results (by echo), probably because of improved surgical technique. In the early part of the series, it was not uncommon to revise an operation before the patient was taken from the operating room. The 8% incidence of intraoperative revision based on echo findings in group 1 was considered somewhat high (see Table 3). However, with improved understanding of problem areas with certain repairs, results improved quickly, and by the time we had reached 200 patients, the need to revise repairs had decreased substantially and has remained relatively stable at 2% to 3%.During this time, we [12] have evolved improvements on repairs of some defects based on echo findings. Although oxygen saturations and green dye curves were obtained early in the series in an effort to determine the sensitivity of the intraoperative echo, it was quickly learned that these less specific methods were less reliable and had no capacity to demonstrate to the surgeon the specific region of difficulty [Z]. Instead we have learned to evaluate the importance of residual defects using a number of criteria that correlate well with outcome. For example, residual ventricular septal defects are thought to be important if (1) the defect can be seen by twodimensional imaging without color flow, (2) substantial turbulence can be seen on both sides of the defect, (3) the amount of turbulence in the right ventricle is substantial and fills a large part of the chamber as opposed to presenting as a small or well-defined jet, (4) turbulence is noted in an area where the surgeon was concerned with the technical repair, or (5) marked turbulence is noted by color-flow mapping in a common trouble spot (Fig 1). The most immediate useful function of postrepair echo

Ann Thorac Surg 1992;5469143

is to help identify those patients who will benefit from revision before they are taken from the operating room. Although it can be claimed that the reduction in the need for this intervention as the series progressed is due to more surgical experience and better surgical skills, we are confident that the synergistic impact that echocardiography has on the evolution of good surgical techniques is enhanced by a surgical team willing to learn this technology so that the learning curves of surgical technique and ability to evaluate echo complement one another. Furthermore, because the system becomes so easy to use (with a little practice) and because it can be potentially useful in any patient, we currently recommend routine use after repair. In a properly equipped operating room, routine evaluation takes less time and can be more helpful than trying to decide which patients should have imaging depending on results of other screening techniques (eg, oxygen saturations and dye curves). As the typical echo takes 2 to 3 minutes, a surgeon can know quickly and with great confidence that the ventricular septal defect is well closed or that a repaired valve is functioning well. In the process of learning techniques to evaluate the presence of residual defects, we found several technical considerations useful: (1) It is important that the transducer make good contact with the epicardial surface. Although this may induce some ectopy, the ectopy is usually well tolerated by the patient and rarely results in hemodynamic compromise. (2) Patients vary with respect to their echogenicity, and occasionally there are patients in whom echo evaluation is more difficult. It is in these patients that experience is useful in enabling the surgical team to interpret information that is less than optimal in quality. Early during a team’s experience, these poorquality examinations must be tolerated, and in time, the information they provide will become more apparent. (3) The use of transesophageal echocardiography in conjunction with epicardial echocardiography is quite helpful. Prosthetic patch material often causes interference with sound wave transmission below the patch, making it difficult to evaluate certain areas after repair [6]. The use of the epicardial and transesophageal approaches concurrently reduces error in interpretation resulting from patch dropout. (4) It is helpful to interrogate the repair briefly during the period of rewarming. This can be done, once the heart has resumed an appropriate rhythm, by temporarily reducing venous return to the pump and by ventilating the lungs. This restores a more normal physiology and enables interrogation of the repair quality. In the occasional instance when an unexpected problem is found with the repair (such as a residual ventricular defect), revision can be performed in an efficient and timely manner without having waited for rewarming and weaning from bypass. Nevertheless, it is important to repeat the examination after the patient has been weaned from bypass. This is especially critical when evaluation requires the more normal loading conditions that exist off cardiopulmonary bypass (such as evaluation of mitral valve function). Recent enthusiasm [13, 141 for the application of newer transesophageal technology to the infant patient after

UNGERLEIDER ET AL ECHO FOR CONGENITAL HEART SURGERY

Ann Thorac Surg 1992;54:691-8

A

C

B

D

695

Fig I. ( A ) Long-axis view of a residual ventricular septa1 defect (VSD) after repair of an atrioventricular canal defect. The broken white arrow demonstrates the location where turbulence will begin when color flow is added. The black arrow demonstrates a discernible residual defect that can be appreciated as a gap between the patch and the aortic annulus and that can be visualized without the addition of color flow. ( B ) Turbulence begins below the patch and extends through the defect to fill the right ventricular cavity. (C) Repair of this area. The pledgets used for the suture repair of this defect can be appreciated, and the defect can no longer be visualized without color flow. ( D ) Although there is still a remaining residual defect, it is quite acceptable. Turbulence does not extend very far into the right ventricular cavity, and there is no major turbulence beloui the patch. (Ao = aorta; LV = left ventricle; RV = right ventricle.)

surgical repair deserves comment. Although transesophageal echocardiography is a useful and effective method, the relative nonparticipation by the surgeon may inhibit full concentration by the surgical team on learning to appreciate echocardiographic images. The epicardial approach gets the surgeon involved and enables him or her to interrogate the quality of the repair with the spatial knowledge of that repair that only the surgeon has. This is not to say that transesophageal echocardiography will not be helpful. It is extremely useful for certain problems (eg, mitral valve function below a prosthetic patch or continuous assessment of ventricular function). However, in the hands of an experienced surgical team, the value and efficacy of epicardial echo in allowing the surgeon to quickly validate the quality of the technical repair should not be minimized. Even as better technology provides

biplane transesophageal echocardiographic probes in infant sizes with reasonable resolution, the same advances will propel corresponding improvements in the epicardial transducers, and once surgeons learn to use the epicardial approach, the transducers become "omniplane." With experience comes an appreciation for those problems disclosed by the postrepair echo that correlate with a bad outcome (ie, death or the need of reoperation). The more recent groups in our series demonstrated excellent correlation of prognosis and postrepair echo findings, suggesting that after approximately 200 intraoperative echo examinations, a surgical team is able to very reliably predict which patients have the highest likelihood of a good outcome based on the postrepair echo findings. Although the terms used to categorize the repairs seem subjective, long-term outcome is objective, and once

696

UNGERLEIDER ET AL ECHO FOR CONGENITAL HEART SURGERY

outcome correlates with the subjective appreciation of the surgical repair (by echo), then the team will know that it has learned to accurately understand the intraoperative information. The most difficult portion of the learning process for the surgical team is to react appropriately to a disturbing postrepair echo finding. It is easy for the team to become overly zealous and attempt to achieve echo-perfect results with each repair. This is clearly unnecessary and not always in the patient’s best interest. Conversely, it is sometimes emotionally difficult for a surgeon to demonstrate a technically less than perfect repair to all members of the operating team and remain feeling secure that the result is truly acceptable. Most agonizing of all are those circumstances where an unexpected but major residual problem is disclosed that requires repeat surgical intervention. These patients frequently have no concomitant clinical problems and can easily be weaned from bypass. The problem can be ignored and the patient can be returned to the intensive care unit in good condition. The agonizing part is that the surgical team learns with time that these patients ”will be back,” and it becomes far easier to do something sooner rather than later [ll].With acquired skill in interpreting the intraoperative echo, the surgical team will find that the surgical procedures will less frequently repeat problems of previous cases. In conclusion, dedicated use of intraoperative echocardiography during the repair of congenital heart defects in 621 patients by the same surgical team demonstrates four important points. (1)Intraoperative echocardiography can be learned by the surgical team to provide information that can prevent patients from leaving the operating room with a suboptimal repair. (2) The surgical learning curve for this technique is approximately 200 cases, but refinements in the ability to use this information continue beyond 600 cases. (3) With experience, postrepair echo can be expected to provide information that will encourage revision of the repair before the patient leaves the operating room in 2% to 3% of cases. The specific information provided by echo enables these revisions to have a high degree of efficiency and success. (4)Once skills are mastered, the ability to correlate intraoperative echo with outcome approaches 97%. ~~

We thank James NeSmith (RDCS) for his technical assistance in the preparation of the manuscript.

Ann Thorac Surg 1992;54:691-8

References 1. Ungerleider RM. The use of intraoperative echocardiography with Doppler color flow imaging in the repair of congenital heart defects. Echocardiography 1990;7:289-304. 2. Ungerleider RM, Greeley WJ, Sheikh KH, Kern FH, Kisslo JA, Sabiston DC Jr. The use of intraoperative echo with Doppler color flow imaging to predict outcome after repair of congenital cardiac defects. Ann Surg 1989;210:526-34. 3. Ungerleider RM, Kisslo JA, Greeley WJ, Van Trigt P, Sabiston DC Jr. Intraoperative prebypaas and postbypass epicardial color flow imaging in the repair of atrioventricular septal defects. J Thorac Cardiovasc Surg 1989;98:90-100. 4. Canter CE, Sekarski DC, Martin TC, Guitierrez FR, Spray TL. Intraoperative evaluation of atrioventricular septal defect repair by color flow mapping echocardiography. Ann Thorac Surg 1989;48:544-50. 5. Gussenhoven EJ, van-Herwerden LA, Roelandt J, Ligtvoet KM, Bos E, Witsenburg M. Inlraoperative two-dimensional echocardiography in congenital heart disease. J Am Coll Cardiol 1987;9:565-72. 6. Hagler DJ, Tajik AJ, Seward JB, Schaff HV, Danielson GK, Puga FJ. Intraoperative two-dimensional Doppler echocardiography. A preliminary study for congenital heart disease. J Thorac Cardiovasc Surg 1988;95:516-22. 7. Sutherland GR, van-Daele ME, Stumper OF, Hess J, Quaegebeur J. Epicardial and transesophageal echocardiography during surgery for congenital heart disease. Int J Card Imaging 1989;4:3740. 8. Dan M, Bonato R, Mazzucco A, et al. Value of transesophageal echocardiography during repair of congenital heart defects. Ann Thorac Surg 1990,50:63743. 9. Hsu YH, Santulli T Jr, Wong AL, Drinkwater D, Laks H, Williams RG. Impact of intraoperative echocardiography on surgical management of congenital heart disease. Am J Cardiol 1991;671279-83. 10. Ungerleider RM. In: Karp RS. Laks H, Wechsler A, eds. Epicardial echocardiography (during repair of congenital heart defects. Chicago: Mosby-Year Book, 1992:285. 11. Ungerleider RM, Greeley WJ, Sheikh KH, et al. Routine use of intraoperative epicardial echocardiography and Doppler color flow imaging to guide ancl evaluate repair of congenital heart lesions. A prospective study. J Thorac Cardiovasc Surg 1990;100:297309. 12. Ungerleider RM. The use of intraoperative epicardial echocardiography with color flow imaging during the repair of complete atrioventricular septal defects. Cardiol Young 1992; 2:5&44. 13. Muhiudeen IA, Roberson DA, Silverman NH, Haas GS, Turley K, Cahalan MK. Intraoperative echocardiography for evaluation of congenital heart defects in infants and children. Anesthesiology 1992;76:165-72. 14. Weintraub RG, Sahn DJ. Pediatric transesophageal echocardiography. Present and future. Anesthesiology 1992;76: 159-60.

DISCUSSION DR ROBERT M. SADE (President): This is a very impressive series, particularly with respect to your ability to correlate changes in outcome over time as experience with this technique increased. To what extent do you believe intraoperative echocardiography was actually responsible for improvement in the overall results? How much of the improvement was due to our ability to better manage children postoperatively and to the honing of your personal surgical skills and the skills of the rest of your team? Because many things change over time, in an

uncontrolled study like this, it may be difficult to assign alterations in outcome to a particular technical innovation such as the use of echocardiography. DR UNGERLEIDER I think you have probably touched on one of the most important features of learning how to use the technology that is available to us, because it is impossible to single out one particular technology that makes all of our changes and improvements possible. I believe it is fair to say that what

Ann Thorac Surg 1992;54:691-8

intraoperative echocardiography is able to do for a surgeon is track that surgeon's progress in learning techniques and probably teach the surgeon how to perform better operations. Some of the results might be due to the improvements in surgical techniques and some, to improvements in learning how to take care of patients postoperatively. Probably there is no clear way to sort out all of these features. What has become clear to us is that if we learn to use the echocardiographic data and if we understand and evaluate these data as objectively as possible, they become a very good predictor for us of how well a patient will do in the postoperative period regardless of all the other factors that might relate to a good outcome. DR RONALD E. ELKINS (Oklahoma City, OK): That was a very nice presentation. My colleagues and I share your enthusiasm for the use of intraoperative echo. The one area that has plagued us is how to define "acceptable" regurgitation through a repaired mitral valve. Could you give us some insight into what you consider acceptable versus unacceptable in the patient who has undergone atrioventricular canal repair? DR UNGERLEIDER Dr Spray, one of the discussants, may be able to answer that question better than I. It has been our impression that almost every patient has some degree of mitral insufficiency discernible by echocardiography after atrioventricular canal repair. In my experience, I have never had to intervene during that particular operation to do anything for that mitral insufficiency. In our series of about 50 patients having atrioventricular canal repair, 2 have had later mitral valve replacement. Neither patient had Down's syndrome, and both left the operating room with very trivial mitral insufficiency. Assessing mitral insufficiency is difficult; it is probably the hardest assessment to correlate with long-term outcome, but the disease itself helps us appreciate that the condition of some patients will deteriorate in terms of mitral valve function in the long term. Conversely, we have seen some patients with moderate to severe mitral regurgitation after repair of atrioventricular canal who get better postoperatively. Perhaps this is due to the change in systemic afterload when they no longer have such a constricted peripheral circulation. In summary, we appreciate the degree of mitral insufficiencyin the operating room but rarely do anything about it. DR CONSTANTINEMAVROUDIS (Chicago, IL): This is a very nice report by Dr Ungerleider and his associates and is years ahead of its time. Unlike most operations, which can be visually assessed after completion, intracardiac repairs cannot be seen and must be evaluated by thrills, pressure tracings, oxygen saturation comparisons, and so on. Routine intraoperative echocardiography represents an excellent technique to evaluate intraoperative repairs, which occasionally need to be revised immediately instead of later. It appears that we may have to learn this technique ourselves for a timely evaluation and not depend on our busy cardiology colleagues. Do you foresee extended inservice requirements for the rest of us to acquire these skills? Also, do you think the transesophageal echocardiogram is a little easier to interpret than the epicardial echocardiogram? DR UNGERLEIDER Thank you for your kind remarks. To answer first the second question regarding the use of transesophageal technology, that technology is improving. Smaller probes with better resolution will become available. They will be inserted by the anesthesiologist and used to provide continuous on-line monitoring of ventricular function and the repair, and this may make it a little less necessary for surgeons to learn these techniques. However, it has been our impression that epicardial echo

UNGERLEIDER ET AL ECHO FOR CONGENITAL HEART SURGERY

697

adds a lot, especially because in the hands of the surgeon who has performed the repair, it can precisely evaluate areas about which the surgeon has concerns. As you know, there are often instances when you have some concern about an operative repair, and with this technology, you can evaluate it yourself. With respect to in-service training, learning this technique is much easier than most other aspects of neonatal heart surgery. All it really takes is a surgeon who is willing to tolerate the learning curve of 200 cases, an echo machine in the operating room, and a cardiologist who is willing to help for the first few cases until you learn how to obtain some of the standard views. Then over a few years you will learn how to interpret the information and you will teach yourself. There is really nothing very difficult about this technology. DR THOMAS L. SPRAY (St. Louis, MO): I compliment Dr Ungerleider for h y n g to bring this technique to the rest of us who do congenital heart surgery. At St. Louis Children's Hospital, we have been using intraoperative echocardiography with both the epicardial and the transesophageal technique for the past 5 years or so but in more selected patients. I have a comment to make on the issue of mitral regurgitation and then a question to ask. I think it is difficult to assess the magnitude of mitral regurgitation because it is so dependent on the hemodynamics at the time you are analyzing it. We have found that sometimes jets of mitral regurgitation that look major by echocardiography turn out not to be very important when contrast is added. We inject agitated blood or a mixture of blood and saline solution, which gives a nice microbubble type of contrast and a more quantitative measurement of the magnitude of mitral insufficiency, directly into the left ventricle. On occasion, this technique has allowed us to accept a degree of mitral insufficiency that by regular color Doppler echo we would have thought too severe and we would have repaired. On several occasions in atrioventricular canal repairs, we have had to go back and put in an extra annuloplasty stitch or a stitch in the commissure between the left superior and inferior leaflet components to control what we consider an unacceptable magnitude of mitral regurgitation. So I think echo assessment is very useful after atrioventricular canal repair. My question is, in how many of your patients with an echo-imperfect result was there a surgical concern at the time? With the large number of patients reported, obviously many did not have an atrioventricular canal or a ventricular septa1defect or some kind of intracardiac shunt lesion, but they may have had conduit placement or repair of the right ventricular outflow tract, which we find very difficult to assess echocardiographicallyby an epicardial technique. In some of these patients, there was minimal concern about your operative repair, and yet echocardiography was used in all of them to see what the repair was like. How many times were you not concerned about your repair and found something on echo that made you want to go back and do something in the operating room? DR UNGERLEIDER That question, Dr Spray, is an excellent one. It is humbling for a surgeon to find out that the results in reality are not always the results perceived at the time the repair was finished. Before we looked at the postrepair echo of the first 328 patients, I was asked by the team, "Well, are you satisfied? Did you do a good job'? And 312 times I said "Yes," and almost 15% of the time, echo disclosed an imperfect repair about which we had some concerns. Whether we did something about it or not, the point is that I was not as happy as I thought I should have been. Sixteen times I told the team, "Gee, that is the best I can do. I am

698

UNGERLEIDER ET AL ECHO FOR CONGENITAL HEART SURGERY

kind of concerned, but I am afraid to try to do more.” Thirty percent of the time the repair turned out to be okay. Clearly, an intraoperative assessment based on the surgeon’s impression of the results is not always accurate, and echo adds a slightly more objective analysis to the repair evaluation. And echocardiography continues to affect our results. Just 2 weeks ago we had a visitor, as always happens when this occurs, and I thought we had done an echo-perfectrepair of an atrioventricular canal, but when I put the probe on the heart, there was a residual ventricular septa1 defect in the inlet septum. So a surgeon’s impression can sometimes be altered by the echo technology. I think echocardiography continues to add information. DR ERIC L. CEITHAML (Jacksonville, FL): Dr Ungerleider, I congratulate you on this presentation. My associates and I also

Ann Thorac Surg 1992;54:691-8

use the epicardial echo technique intraoperatively to assess congenital heart repairs, but we have had a difficult time visualizing the right ventricular outflow tract, particularly after homograft reconstruction in patients with tetralogy or pulmonary atresia. Do you use any particular technique to visualize this area? DR UNGERLEIDER It is possible to look at the right ventricular outflow tract at least with respect to insufficiency. However, I think it is a misapplication of echo to use it to get an idea of pressure measurements. It is easy for the surgeon to simply insert a needle into various portions of the repair to find out where the pressure gradients, if any, exist. So I think it is wrong to try to make echo something it is not. For right ventricular outflow repair, we recommend pressure measurements.